Project 1. Azole resistance in Candida glabrata. A major and long term project of the laboratory has been to understand why azole resistance arises rapidly during treatment of Candida glabrata infections in humans, in contrast to other pathogenic yeasts. Transcriptional control of azole resistance: a. PDR1p is the major transcriptional factor controlling azole efflux and drug resistance in Candida glabrata. This protein has been postulated to bind to the promoter region of many target genes but the gene targets and their upstream binding sites have not been defined clearly. We are using ChIP-chip for a genome-wide screen of PDRIp DNA binding sites. We have thus far confirmed the binding of PDR1p to two previously postulated sites and are exploring the numerous other targets identified by this technique. Although we are using a His-6 tag on PDR1p for ChIP-chip analysis, it would preferable to have antibody against this protein. CgPdr1p contains 1107 amino acids. In order to produce a recombinant protein for antibody production, the CgPDR1 gene was divided into three segments and the three protein fragments overexpressed in E. coli.. Expression of the three partial genes was confirmed but the N terminus had the best expression among the three. Large scale protein purification is in progress and the purified protein will be used for antibody production in rabbits. b. Transcriptomes of azole susceptible and resistant isolates. In a continuation of work reported last year with quantitative RT-PCR, we used microarray to analyze genome-wide changes in the transcriptome of isolate pairs. Transcriptional profiling revealed that a cluster of 15 genes was upregulated in the majority of azole resistant isolates when compared to their paired susceptible isolates. These 15 genes were categorized into four major biological processes, which included response to chemical stimuli, transport, response to stress, and transcription. This co-regulation indicates a shared pathway of azole resistance in clinical isolates. c. Study of CgSTB5 and its role in Candida glabrata azole susceptibility In prior studies we have shown that one of the azole resistance mechanisms was gain-of function mutations in the transcriptional regulator gene, CgPDR1. The gene upregulates the expression of multidrug resistance genes, i.e. CgCDR1 and PDH1. In Saccharomyces cerevisiae, Stb5p was reported to form a heterodimer with Pdr1p. Stb5p, like Pdr1p, contains a Zn(II)2Cys6 zinc finger domain that interacts with a pleiotropic drug resistance element in vitro. Stb5p also regulates oxidative stress response. The S. cerevisiae stb5 null mutant has been reported to exhibit reduced vegetative growth as well as increased sensitivity to caffeine, cold, oxidative stress, and cycloheximide, facts which we confirmed. However, deletion of stb5 did not affect its sensitivity to azoles. During this past year we found that deletion of the homologous gene in C. glabrata, CgSTB5, in both the wildtype strain and in a pdr1 null mutant increased resistance to azoles. Additionally, overexpression of CgSTB5 in C. glabrata decreased resistance to azoles, suggesting that this gene is a transcriptional repressor rather than an activator. But the transcriptional targets are none of those anticipated, judging by unaltered expression of CgCDR1 and CgSNQ2 . Also unlike S. cerevisiae, deletion of CgSTB5 in C. glabrata did not alter sensitivity to caffeine, cold, or oxidative stress. However, we found that CgSTB5 was a functional homolog of STB5 by genetic complementation analyses. CgSTB5 was able to restore the stb5 null mutants defects in vegetative growth, caffeine and cold sensitivity. The increased fluconazole susceptibility associated with the overexpression of CgSTB5 is unlike previously described mechanisms and will be useful to pursue. Project 2. Voriconazole metabolism, genotype and toxicity. Voriconazole is a major drug used for treatment of fungal infections in immunosuppressed patients. Wide variability in blood levels has been noted and correlated in Japanese patients with a specific 2C29 genotype. This gene product is thought to control voriconazole conversion to the N-oxide, a metabolite with no known biologic activity. It is unclear whether this genotype confers the same phenotype in other racial backgrounds. High voriconazole blood levels have been said to predict toxicity but the information is too fragmentary to be useful. In our ongoing study of voriconazole metabolism and genotype, since the last annual report we have increased the number of voriconazole trough serum assays from 171 to 269, and increased from 62 to 84 the number of patients under study. We have also increased the number of patients to 79 in whom we have sequenced the 2C19 gene and its -806 promoter and sequenced the 2C9 gene in 69 patients. Toxicity was monitored prospectively and found to include visual changes in 19%, hallucinations 18%, photosensitivity 12%, and hepatotoxicity in 8%. Trough plasma levels were higher only in patients with hallucinations and not with visual changes, photosensitivity or hepatotoxicity. A new finding of concern is that the voriconazole level was below the limit of our assay (0.2 mcg/ml) in 17 specimens from 11 patients. The N-oxide metabolite was higher than 0.2 mcg/ml in all but one patient, indicating that metabolism and not noncompliance was the likely explanation for low voriconazole serum levels. Rather, use of a low oral dose (200 mg twice daily) accounted for the low concentration in 14 of 17 specimens, indicating that this oral dose may be inappropriately low for adults. As we reported last year, the 2C19 slow metabolizer genotype described in Japanese continued not be correlated with toxicity or plasma levels of voriconazole or its metabolites in our patients. The 2C9 genotype also was not correlated with toxicity or metabolism. Project 3. Gene expression of C.glabrata within the spleen of infected mice. We continued the project described last year, in which gene expression of C. glabrata was studied during phagocytosis by human neutrophils. This year we used qRT-PCR to analyze gene expression in the spleens of infected C57BL/6J mice and gp91phox-/- knockout mice from the same background. Phagocytes of these knockout mice have defective oxidative microbicidal activity. We have previously reported that these knockout mice permitted rapid growth of C. glabrata in tissue, leading to death of the mice. In contrast, wild type mice of the same strain cleared the fungus after infection and survived. First, we examined the change in expression of 20 C. glabrata target genes chosen to reflect major metabolic pathways. We studied C. glabrata gene expression in spleens of infected C57BL/6J mice and compared that with expression during phagocytosis by neutrophils. The results showed that expression of 15 genes were similar whereas 3 genes (AUS1, PCK1, CDC19) were significantly upregulated more than eight fold in mouse spleens;the other two genes had lesser levels of upregulation. These results suggested C. glabrata cells in mouse spleens were responding to more severe carbohydrate depletion (PCK1, CDC19) and were attempting to uptake cholesterol (AUS1), compared to conditions within neutrophils. We compared the results in spleens of infected wild type mice to those in spleens from infected gp91phox-/- mice. Transcriptional evidence of stress on C. glabrata was much less in the knockout mice. In comparison, the wild type mice showed remarkable upregulation of MET28 and other genes involved in methionine metabolism as well as genes controlling alternative carbohydrate sources, MAP kinase, sterol transporters (AUS1) and xenobiotic transporters (PDR1).